Contents lists available atScienceDirect
Applied Geochemistry
journal homepage:www.elsevier.com/locate/apgeochem
The Campo de Calatrava Volcanic Field (central Spain): Fluid geochemistry
in a CO2-rich area
B. Nisi
a,∗, O. Vaselli
b,c, J. Elio
d, L. Giannini
b,c, F. Tassi
b,c, M. Guidi
a, T.H. Darrah
e, E.L. Maletic
e,
A. Delgado Huertas
f, S. Marchionni
baCNR- IGG (Institute of Geosciences and Earth Resources), G. Moruzzi, 1, 56124, Pisa, Italy bDepartment of Earth Sciences, Via G. La Pira, 4, 50121, Florence, Italy
cCNR- IGG (Institute of Geosciences and Earth Resources), G. La Pira, 4, 50121, Florence, Italy
dETS Ingenieros de Minas Universidad Politécnica de Madrid, Calle de Ríos Rosas, 21, 28003, Madrid, Spain
eSchool of Earth Sciences, Ohio State University, 275 Mendenhall Laboratory, 125 South Oval Mall, Columbus, OH, 43210-1398, USA fInstituto Andaluz de Ciencias de la Tierra, CSIC-UGR, Avda. de las Palmeras 4, Armilla, Spain
A R T I C L E I N F O Editorial handling by Dr M Liotta
Keywords:
Campo de Calatrava volcanic field Central southern Spain Mg-HCO3waters
Geopressurized CO2-Rich reservoir
A B S T R A C T
The Campo de Calatrava Volcanic Field (CCVF) located in central-southern Spain (along with Selva-Emporda in Catalonia, NE Spain) is regarded as one of the most important CO2emitting zones in Peninsular Spain. Here, we report and evaluate new molecular and isotopic geochemistry of thermal waters and CO2-rich gas discharges from the CCVF. Locally, these CO2-rich fluid emissions represent the remnants of the past volcanic activity that affected this area from the late Miocene through the Quaternary, with the most recent events occurring in the Holocene. The locations of discharging fluids and previous volcanic centers appear to be aligned along well-defined NW-SE and NNW-SSE lineaments, with subordinate trends in the ENE-WSW direction. The chemical and isotopic composition of the thermal waters suggests a meteoric origin, dominated by three distinct geochemical facies: 1) HCO3-Mg(Ca) type waters, associated with a relatively shallow aquifer and related to the interaction of meteoric waters with CO2-rich gases, alkaline volcanic products, and sedimentary formations, 2) SO4(Cl)-Ca(Mg) type waters, which stems from the two rivers (Guadiana and Jabalón) that drain Triassic evaporitic rocks before entering the study area, and 3) HCO3-Na type waters, hosted in deep geopressurized CO2-rich reservoirs within the Ordovician basement rocks.
The87Sr/86Sr isotopic compositions (ranging between 0.70415 and 0.71623) and δ34S-SO
4values (+10.7 to +18.3‰ vs. CDT) of CO2-rich fluids are consistent with interactions between water and either the Paleozoic basement, Triassic evaporites, Quaternary volcanic rocks, or a combination thereof. Dissolution of a CO2-rich gas phase into the aquifer produces low pH values (down to 5.4) and enhances water-rock interactions causing relatively high salinity (Total Ionic Salinity: up to ∼185 meq/L). Carbon dioxide is by far the most abundant gas constituent (up to 992 mmol/mol) and is dominated by mantle-derived sources as indicated by the combination of relatively high helium isotopic ratios (up to 2.7 R/Ra), high isotopic ratios of carbon in CO2(ranging between −6.8 and −3.2‰ V-PDB), and the carbon isotopic signature of TDIC (from −6.8 to +2.2‰ vs. VPDB).
In the last two decades, numerous (CO2-rich) gas blowouts have occurred in the area during well drillings, suggesting the presence of a geopressurized gas reservoir at relatively shallow depth.
1. Introduction
There are four areas of Neogene volcanic activity in continental Spain, including the Internal Betics (SE Spain; 34–2 Ma), the Valencia Trough (24–0.01 Ma), the Calatrava Volcanic Province (9–0.7 Ma), and Olot–Garrotxa (10–0.01 Ma). A fifth area of Late Miocene (12.1-6.1 Ma) volcanic activity is present in the Alboran Basin (westernmost
Mediterranean Sea between the southern coast of Spain and northern Morocco) where dacites, rhyolites, and granites occur with significant volumes of tholeiitic to calc-alkaline basalts, basaltic andesites, and andesites (e.g. Duggen et al., 2004, 2005 and references therein) (Fig. 1).
Out of these five volcanic fields, three areas exhibit CO2-rich gas emissions. The majority of CO2 emissions is associated with
low-https://doi.org/10.1016/j.apgeochem.2019.01.011
Received 2 October 2018; Received in revised form 22 January 2019; Accepted 24 January 2019
∗Corresponding author. CNR-IGG Institute of Geosciences and Earth Resources, Via Moruzzi, 1, 56124, Pisa, Italy. E-mail address:b.nisi@igg.cnr.it(B. Nisi).
Applied Geochemistry 102 (2019) 153–170
Available online 31 January 2019
0883-2927/ © 2019 Elsevier Ltd. All rights reserved.
thermal waters and occurs in areas with recent volcanic activity. Prominent CO2gas discharges are recognized at: 1) Olot–Garrotxa (in the NE Volcanic Province, hereafter NEVP), 2) the Internal Betics (in the SE Volcanic Province, hereafter, SEVP), and 3) Campo de Calatrava (south-central Spain, Volcanic Field, hereafter, CCVF) (Fig. 1A). These volcanic provinces are likely part of an aborted rift, which is most clearly expressed in the Rhine Valley, and lie along the Trans-Moroccan Western Mediterranean-European Fault Zone (TMWMEFZ,Fig. 1A; e.g. López-Ruiz et al., 2002). While the NEVP and CCVF are dominantly characterized by alkaline basaltic volcanism (e.g. Cebriá and López-Ruiz, 1995;Cebriá et al., 2000), the SEVP displays calc-alkaline, high-K calc-alkaline, shoshonitic, ultrapotassic, and alkaline basaltic volcanics
(Duggen et al., 2005; Cebriá et al., 2009). In the CCVF, minor Late-Miocenic volcanic episodes (8.7–6.4 Ma) of leucititic eruptions were followed by alkali basalts, basanites, ol-nephelinites, melitites, and carbonatites in the Pliocene-Quaternary during which diatremes and maars formed (López-Ruiz et al., 1993;Ancochea, 2004;Bailey et al., 2005). Most volcanic deposits contain mantle xenoliths sourced from depths of > 70 km (e.g.Bianchini et al., 2010;Martelli et al., 2011).
Most CCVF volcanoes intruded the Palaeozoic basement in the Calatrava and Almagro massifs, the latter showing E-W and N-S ver-tical, flexural folds (De Vicente et al., 2007). Such massifs are affected by NW-SE and E-W-oriented fault systems to produce a horst and graben morphology (e.g.Stoppa et al., 2012), along which the main
Fig. 1. a) Location of the Trans-Moroccan, Western Mediterranean, European Fault Zone (TMWMEFZ,López Ruiz et al., 2002), the Campo de Calatrava Volcanic Field (CCVF), and the Western Mediterranean European Block (WMEB); b) Schematic geological map of the Iberian Peninsula and c) Map of the Calatrava Volcanic Field with the locations of sampled waters (by©National Geographic Institute: IGN;http://www.ign.es/wms-inspire/pnoa-ma).
CO2-rich gas emissions are found (Poblete Piedrabuena, 1997;González Cárdenas and Gosálvez Rey, 2004).
In this work, we examine thermal waters and gas discharges from the Campo de Calatrava Volcanic Field, where in the last few decades the presence of a CO2-pressurized reservoir at a relatively shallow depth in this region has caused several small-sized explosions or gas blowouts, particularly during the drilling (down to 200 m) of domestic water wells (e.g. González Cárdenas et al., 2015). The main objectives of this in-vestigation are to i) describe the geochemical and isotopic features of the thermal water and gas discharges in the CCVF from samples col-lected in July 2009 and July 2012; ii) determine if the root of mantle-derived volcanism that led to the formation of the CCVF is still present and actively communicating with the surface, and iii) provide a con-ceptual model based on the available geochemical and isotopic data. 2. Geological and volcanological setting
During the last 60–70 million years, the areas in and around the Mediterranean Sea have experienced extensive igneous activity (e.g. Wilson and Downes, 1991). The Campo de Calatrava Volcanic Field (CCVF) is one of the main Cenozoic magmatic provinces belonging to the circum-Mediterranean region (Lustrino and Wilson, 2007). The volcanic activity in this area has variably been attributed to either small mantle plume/hot spots (Cebriá and López Ruiz, 1995; Wilson and Patterson, 2001; Bell et al., 2013) or the complex Trans-Moroccan Western Mediterranean-European Fault Zone (TMWMEFZ,López Ruiz et al., 2002), the latter likely being generated by a large asymmetric mantle upwelling related to the former North South America-Africa triple junction (Oyarzun et al., 1997).
From the Miocene to present, alkaline basaltic volcanism occurred within and is likely associated with TMWMEFZ (López Ruiz et al., 2002). The TMWMEFZ offers a peculiar tectono-magmatic scenario that extends from the northern Europe through the Alpine arc to the Pyr-enees (López Ruiz et al., 2002;Doblas et al., 2007) (Fig. 1A) and bounds the west-directed Western Mediterranean European Block (WMEB; López Ruiz et al., 2002). The current understanding suggests that CCVF is an expression of lateral extension of the TMWMEFZ along the edge of the Central Iberian Zone of the Iberian Massif, close to the outer sectors of the Alpine Betic Range (Fig. 1B).
The geology of the study area includes Paleozoic basement rocks covered by late Cenozoic sediments. The Paleozoic rocks are mainly quartzites belonging to the so-called Armorican facies (lower Ordovician), which are overlain by slates, sandstone interbeds, and discontinuous carbonate (upper Ordovician) deposits that are folded from NW-SE to W-E (López-Ruiz et al., 1993;Gutiérrez-Marco et al., 2002). Extensive outcrops of Triassic rocks occur in the eastern portion of the CCVF, including the Germanic facies (Buntsandstein, Mu-schelkalk, and Keuper). Upper Miocene to Quaternary fluvial and la-custrine sediments were deposited within fault-bounded Tertiary-Qua-ternary basins related to the Late Miocene extensional tectonic activity, and unconformably overlie the basement rocks (Fig. 1C) (Ancochea and Brändle, 1982;López Ruiz et al., 1993,2002;Cebriá and López Ruiz, 1995;Carracedo Sánchez et al., 2012;Herrero-Hernández et al., 2015). The CCVF was the primary location for late Miocene-Quaternary volcanic activity in central Spain (Fig. 1C). Vents and outcrops of mafic lava flows and pyroclastic deposits of alkaline composition are scat-tered throughout the area, which cover approximately 5000 km2(e.g., Ancochea, 1999;Gonzalez Cardenás et al., 2007;Stoppa et al., 2012). The CCVF volcanic rocks are part of an intracontinental plate magmatic association (Cebria and Lopez-Ruiz, 1995;López-Ruiz et al., 1993; 2002) consisting of mafic silica-undersaturated alkaline lavas (from alkali basalts and nephelinites to melilitites and leucitites) (Cebrià and Lopez- Ruiz, 1995;López-Ruiz et al., 1993,2002;Stoppa et al., 2012) which commonly host mantle xenoliths (Martelli et al., 2011 and references therein) and carbonatites (Bailey et al., 2005, Humphreys et al., 2010; Stoppa et al., 2012). The CCVF volcanics
display a close affinity to the Miocene-Quaternary volcanic regions of western and central Europe (Wilson and Downes, 1991). Trace element geochemistry displays enrichments of incompatible elements that were interpreted as a mixture of lithospheric and asthenospheric mantle (with affinities to a HIMU-OIB source mantle), which experienced a small degree of partial melting (e.g. Cebrià and Lopez-Ruiz, 1995; Martelli et al., 2011).
The main eruptive features that characterized the CCVF were dominantly caused by strombolian-type and hydromagmatic eruption events (Ancochea, 1999;Gonzalez Cardenás et al., 2007,2010;Stoppa et al., 2012;Becerra-Ramírez et al., 2010;Stoppa F. and Schiazza M., 2013) without any reported evidence of hawaiian-type eruptions (Ancochea, 1999), contradicting what was reported by Carracedo Sànchez et al. (2009). The hydromagmatic eruptions play a key role in the volcanological history of CCVF, since most volcanic centers show deposits related to interactions between magma and water. Occasion-ally, hydromagmatic products alternate with those related to strom-bolian-type eruptive activity. The crater bottoms have often developed endorheic or subendorheic areas, where small evaporitic deposits are also found (Ancochea, 1999). The CCVF can be subdivided into two phases (Ancochea, 1982,1999) based on the age and composition of the volcanic products. The first phase, which consists of ultrapotassic vol-canics is less intense and occupies the central part of the region. Radiometric ages suggest that the emplacement of these volcanics oc-curred between 8.7 and 6.4 Ma. The second phase, which includes al-kaline and ultra-alal-kaline volcanics, was emplaced between 3.7 and 0.7 Ma. The best preserved volcanic edifices are synchronous and suc-cessive to the detritic-carbonate deposit of the Upper Pliocene (Portero et al., 1984). In the central part of the region (Ciudad Real and Al-magro) the Plio-Pleistocene erosional surface is intruded and deformed by several volcanic centers. Note thatGonzáles Cárdenas et al. (2007) attributed an age of 5550 BP to Columba volcano, which is located about 8 km south of La Sima (Fig. 1C).
The CCVF is characterized by a complex fracture pattern that con-trols the geometries of the E-W to ENE-WSW, NW-SE, and NE-SW basins (Crespo, 1992), as well as the facies and thickness of the Cenozoic continental sediments. In addition to the Late Miocene extensional phase, at least two Neogene tectonic episodes in the CCVF took place: (i) the opening of the La Mancha Basin and (ii) a weak regional-scale compressional phase (IGME, 1988). Consequently, the region shows a basin and range-like morphology. The ranges are relatively high with elevations between 700 and 900 m and highlands with an elevation of approximately 600 m. The Guadiana River and the Jabalón River, which is the main tributary to the Guadiana River in its upper reaches, are the two main rivers that drain the study area. Both rivers have a roughly NS and NW–SE-orientation, respectively (Poblete Piedrabuena et al., 2016) and interact with extensive evaporitic outcrops of Triassic age before flowing through the CCVF.
3. Thermal waters and CO2-rich emissions
The CCVF has a complex and intriguing volcanic and tectonic set-ting, which in combination with that of Selva-Empordá, accounts for the majority of natural CO2 emissions within Peninsular Spain (Catalonia; e.g. Vaselli et al., 2013; Elio et al., 2015 and references therein). As mentioned, the CCVF is a relatively young volcanic field and hosts a large number of springs. Both are regarded as an economic/ touristic resource for the region (Escobar and González, 2010; Becerra-Ramírez et al., 2017).
The thermal springs are often accompanied by CO2-rich gas bub-bling pools, locally known as “hervideros” (e.g.Yélamos et al., 1999; Melero Cabañas, 2007), and CO2-rich dry gas vents (e.g.Peréz et al., 1996;Melero Cabañas, 2007;Vaselli et al., 2011,2012). The manifes-tations of the CCVF are primarily located in the southern sub-plateau of the Castilla-La Mancha region and are aligned with well-defined lineaments that trend NW-SE, NNW-SSE, and subordinately ENE-WSW
(Melero Cabañas, 2007). Locally, the surface CO2-rich emissions are mainly associated within zones of more intense fracture systems and are commonly found as small (< 1 m2) emission sites. The gas emission rate ranges from barely observable gentle bubbling to intense and vigorous fluxes of CO2. Among the CO2emissions in the area, La Sima (dry gas vents,Fig. 2a), Cañada Real (degassing pools,Fig. 2b), and Jabalón River (degassing pools and springs,Fig. 2c) can be considered the three most representative sites of the CO2seepage in the CCVF.
La Sima is a CO2-rich gas discharge (up to 2 tons of CO2per day) that emits from a restricted surface depression of approximately 5 m in diameter (Fig. 2a) (Elio et al., 2015), where numerous small dead an-imals such as lizards, mice, and birds are found. The Cañada Real dis-charge (Vaselli et al., 2013;Gasparini et al., 2016) is located in the wine yard of the homonymous farmhouse (Municipality of Pozuelo de Ca-latrava). The Cañada Real consists of two pools: the smaller one is about 2 m wide while the larger one is approximately 12 m in diameter (Fig. 2b). The larger pool is characterized by intense gas bubbling with an estimated CO2emission rate between 5 and 20 tons/day, with a mean value of 10 tons/day (Vaselli et al., 2012). The Jabalón River discharge is located 6 km from the village of Granátula de Calatrava (Fig. 2c), where moderate to large CO2-rich emissions (estimated dis-charge rate: 10 tons/day,Vaselli et al., 2012) bubble into < 30°C wa-ters, although the highest recorded temperature in the area is the spa at Baños de Fuencaliente (38°C; e.g.Poblete Piedrabuena, 1992) whose access was prevented by the owners. Most thermal discharges are aligned along a NW–SE fault system that runs parallel to the Jabalón River.
Remarkably, numerous high-pressure gas discharge blasts
(commonly termed blowouts) have been reported to occur in the CCVF due to water well drilling in the region. One of the most famous oc-currences is the “El Chorro” geyser in the Granátula-Moral de Calatrava (Fig. 1C) that discharged from a 200 m deep well and blasted a column of water and gas up to 60 m into the area during drilling operations in 2000 (González Cárdenas et al., 2015). The most recent significant event took place in 2011 (close to the Yesosa volcano, Almagro, Fig. 1C). This discharge, named the “geyser” of Bolaños de Calatrava, spontaneously appeared in a vineyard and produced approximately 50,000 m3of water propelled by gases that covered an area of about 90,000 m2 and discharged up to 40 tons of carbon/day as CO
2 for
several days (Vaselli et al., 2012; Stoppa et al., 2012; González Cárdenas et al., 2015). Minor “eruptive” events were recorded to the SW and E of Almagro at Aldea del Rey and Calzada de Calatrava and Hoya del Peral, El Barranco and Lo Oscuro (in 2011), and El Prado (in 2013), respectively (González Cárdenas et al., 2015). Furthermore, after the seismic crisis that occurred in 2007, a significant increase in gas emission rates (from 0.03 up to 324 kg/m2/day) was registered at La Sima where new CO2-rich gas vents also opened (González Cárdenas et al., 2007;Peinado et al., 2009).
These events are strongly indicative of a geologically pressurized (CO2-rich) reservoir (estimated at about 63 bars during the Bolaños de Calatrava gas blast) that occurred at shallow depths in this region (González Cárdenas et al., 2015). The occurrence of spontaenous gas discharges/blowouts suggests that this area is "overpressured," meaning naturally occurring geological pressures exceed the anticipated hydro-static pressure gradient.
Fig. 2. Selected photos of some of the gas and thermal discharges from the CCVF and the respective I.D. as reported inTables 1–4: a) La Sima; b) Cañada Real; c) Javalon; d) Fuente Gallega; e) El Chorillo; f) El Baño Chico; g) Los Baños de Villa Franca; h) Balneares Cervantes; i) Baño del Trujillo.
4. Materials and methods
4.1. Sampling
Between July 2009 and July 2012, 28 samples were collected throughout the Campo de Calatrava Volcanic Field that included thermal and cold waters, surface waters, dissolved gas samples, and free gas samples. The sampling site locations are shown inFig. 1C. Samples were collected from (i) 7 bubbling pools (#H1, #H3, #H3b, #H7, #VP1, #BDS4, and #FO8); (ii) 9 spring discharges (#H2, #FAG5, #VC6, #CH7, #UBA1, #FO1, #BE1, #FD10, and #FLT9); (iii) 3 groundwater upwelling (#BLC1, #BLC2, and #BLC3); (iv) 3 gas-rich springs (#H4, #H5, and #FAP2); (v) 3 rivers (#H8, #CRE12, and #CRE15); (vi) 1 river with bubbling gas (#JA1); (vii) 1 water well with bubbling gas (#H6), and (viii) 1 dry vent (#LS3). Water and the as-sociated gas from #H1 (Cañada Real) were sampled and analyzed twice (July 2009 and May 2012). During the first sampling, water and gas were collected from the small pool since the big pool at that time was not accessible. Consequently, the big pool was collected during the second sampling trip. Selected images of the dry vents of La Sima (#LS3), the bubbling pool of Cañada Real (#H1), the Jabalón River (#JA1), Fuente Gallega (#H2), El Chorillo and El Baño Chico (#H3 and #H4, respectively), Los Baños de Villa Franca and Balneares Cervante (#H5 and #H6, respectively), and El Baño del Trujillo (#H7) are re-ported inFig. 2a-i.
4.2. Chemical and isotopic analysis of water samples
Temperature, pH, electric conductivity, and alkalinity (titration with 0.01 N HCl and methyl-orange as the indicator) were determined in the field following standard methods. Water samples were filtered (0.45 μm) and stored in high-density polyethylene bottles for laboratory analyses. Cations (Ca2+, Mg2+, Na+, and K+) and anions (Cl−, SO
42−,
F−, Br−, and NO
3−) were analyzed by using AAS (AAnalyst 100 Perkin Elmer) and Ion Chromatography (Dionex 100) on filtered and acidified (0.5 mL Suprapur HCl were added to 50 mL of water) bottles for cations and on filtered only samples for anions, respectively, following methods reported previously (Cuoco et al., 2013;Nisi et al., 2013a). Ammonia (NH4+) was analyzed by molecular spectrophotometry following methods reported previously (Hach DR2100). Trace elements (Al, As, Ba, B, Cs, Co, Cr, Cu, Fe, Hg, Li, Mn, Rb, Se, Sr, and Zn) were de-termined on filtered and acidified (0.5 mL Suprapur HNO3were added to 50 mL of water) samples by ICP-MS by methods reprorted previously. The analytical error for major and trace compounds was < 5 and 10%, respectively.
Four aliquots for each water were sampled for the isotopic analysis, as follows: (i) 125 mL for oxygen and hydrogen in H2O; (ii) 50 mL for carbon in TDIC (Total Dissolved Inorganic Carbon) after adding ap-proximately 2 mg of HgCl2to inhibit carbon isotopic fractionation by bacteria (Atekwana and Krishnamurthy, 1998); (iii) 125 mL with bot-tles pre-cleaned with sub-boiled HCl 6N for Sr dissolved concentrations and87Sr/86Sr isotopic ratios; and (iv) 500 mL for sulfur isotopes in SO
4.
The18O/16O and2H/1H isotopic ratios (expressed as δ18O and δD ‰ vs. VSMOW) were determined by using a Finnigan MAT 250 Delta-S mass spectrometer using standard procedures (Doveri and Mussi, 2014). The analytical precision was 0.1‰ for δ18O and 1‰ for δD. Carbon isotopes in TDIC (expressed as δ13C‰ vs. VPDB) were per-formed with a Finnigan Delta Plus XL mass spectrometer on the CO2 recovered after the reaction of about 3 mL of water with 2 mL of an-hydrous H3PO4in 12 mL pre-evacuated vials (Salata et al., 2000). The recovered CO2 was analyzed after extraction and purification proce-dures on the gas mixture were performed using liquid N2and a solid-liquid mixture of solid-liquid N2and trichloroethylene (e.g. Vaselli et al., 2006,2009). The analytical error for δ13C-TDIC was ± 0.05.
The34S/32S ratios of SO
42−(expressed as δ34S-SO4‰ vs. V-CDT) for seven selected water samples were analyzed using an EA-IRMS (Europa
Scientific, Crewe, UK), equipped with an elemental analyzer (Sercon Ltd., Crewe, UK), after the precipitation of BaSO4 with BaCl2. After centrifugation and drying, the solid phase was transferred into tin capsules with a V2O5catalyst. The capsules were loaded in sequence into a furnace at 1080°C using an automatic sampler and combusted in the presence of O2. Next, the temperature was increased to 1700°C. The combusted gases were then swept in a helium stream over combustion catalysts (tungsten oxide/zirconium oxide) and through a reduction stage of high purity copper wires to produce SO2, N2, CO2, and water. Water was removed using a Nafion™ membrane and SO2was resolved from N2and CO2on a packed GC column at 45°C. The resultant SO2 peak entered the ion source of the IRMS. Gas species of different mass were separated in a magnetic field and simultaneously measured on a Faraday cup universal collector array. Analysis was based on mon-itoring of m/z 48, 49, and 50 of SO+produced from SO
2in the ion source. Reference standards (IA-R025, IA-R026, and IA-R061) were used for calibration and correction of the18O contribution to the SO+ ion beam while working standards were NBS-127, IAEA-SO-5, and IAEA-S-1. The analytical uncertainly was ± 0.3‰.
The 87Sr/86Sr isotopic ratios of six selected water samples were measured in a dynamic mode with a nine-collector Finnigan Triton-TI mass spectrometer (Avanzinelli et al., 2005) using the procedures de-scribed inNisi et al. (2008). External precision of NIST SRM987 in-ternational reference sample for period of this study was 87Sr/86Sr = 0.710246 ± 0.000005 (2σ, n = 40), while the long-term mean value was 0.710248 ± 0.000015 (2σ, n = 186).
Finally, pCO2, saturation index (SI), and TDIC were computed by means of the EQ3 code (Wolery and Jarek, 2003).
4.3. Chemical and isotopic analysis of dissolved and free gas samples
The dissolved gases were collected in pre-evacuated 250 mL glass flasks tapped with Teflon stopcocks according to the procedure re-ported inTassi et al. (2008,2009). The determination of dissolved gases was carried out at equilibrium conditions (STP, Standard Temperature and Pressure). The dissolved inorganic gases in the headspace of the sampling flasks (CO2, N2, Ar, O2, Ne, He, and H2) were measured by gas chromatography with a Shimadzu 15A equipped with a 5 m long stainless-steel column packed with Porapak 80/100 mesh and a Thermal Conductivity Detector (TCD), with the exception of Ar and O2 since a Thermo Focus gas chromatograph equipped with a 30 m long capillary molecular sieve column was used. CH4was analyzed using a Shimadzu 14A equipped with a 10 m long stainless-steel column packed with Chromosorb PAW 80/100 mesh coated with 23% SP 1700 and a Flame Ionization Detector (FID) (e.g.Vaselli et al., 2006;Tassi et al., 2018). The analytical error for GC analysis was ≤5%. The gas species in the liquid phase were calculated according to the Henry's Law constants (Wilhelm et al., 1977).
Gas samples from bubbling pools were collected using pre-weighted and pre-evacuated 50 mL thorion-tapped glass tubes, partially filled with 20 mL of 4N NaOH connected to a plastic funnel positioned over the rising bubbles (Montegrossi et al., 2001;Vaselli et al., 2006). Acidic gases (e.g. CO2and H2S) dissolved in the alkaline solution where the residual gases enriched in the glass tube headspace. The chemical gas composition in the gas vial headspace was determined by gas-chro-matography, while gas concentrations in the liquid phase were de-termined by ion-chromatography (Tassi et al., 2004,2009;Vaselli et al., 2006). A second group of pre-evacuated 50 mL thorion-tapped glass tubes containing 20 mL of 4N NaOH were collected for helium isotope analysis. A pre-evacuated 50 mL gas tube was used to sample the gases for the determination of carbon isotopes in CO2. The La Sima dry vent was collected by inserting a titanium tube into the gas discharge, which was connected to a Tygon™ tube by means of a Dewar glass and then to the sampling vial. The gas sample was collected as previously de-scribed.
TCD-equipped gas-chromatographs (Shimadzu 15a and Thermo Focus). Methane and light hydrocarbons were analyzed with a Shimadzu 14a gas-chromatograph equipped with a FID. Carbon monoxide was de-termined with the same apparatus described for hydrocarbon analysis after its conversion to CH4at 400°C by using a Shimadzu MTN-1 me-thanizer. Analytical precision was < 1% for major gas components and < 5% for minor and trace compounds.
The13C/12C ratios in CO
2were determined by mass spectrometry by using a Finnigan Delta S after a two-step extraction and purification procedure as previously described for the determination of δ13C
TDIC values. The analytical error was ± 0.05‰. The δ13C values of dissolved CO2were calculated from the measured carbon isotopic ratios (δ13
C-CO2strip) in the dissolved gases on the basis of the enrichment factor (ε1)
for gas-water isotope equilibrium proposed byZhang et al. (1995), as follows:
ε1= δ13C-CO2─δ13C-CO2strip= (0.0049 × T) −1.31 (1)
Finally, the elemental abundance of helium (He), neon (Ne), and argon (Ar), and the isotopic analyses of helium (reported as R/Ra where R is the measured helium (3He/4He) isotopic ratio and R
Ais that of the air: 1.384 × 10−6 and corrected using 4He/20Ne ratio), neon, and argon were performed using a Thermo Fisher Helix SFT mass spectro-meter at The Ohio State University Noble Gas Laboratory (OSU NGL), following standard procedures summarized previously (Darrah et al., 2012,2015;Eymold et al., 2018). The average external precision for noble gas concentrations based on “known-unknown" standards was within ± 1.64%, with values reported in parentheses: 4He concentra-tions (0.67%), 22Ne concentrations (1.23%), and36Ar concentrations (0.31%). Noble gas isotopic standard errors were approximately ± 0.0091 times the ratio of air (1.384 × 10−6) for the3He/4He ratio, less than ± 0.371% and ± 0.478% for 20Ne/22Ne and 21Ne/22Ne ratios, respectively, and less than ± 0.224% and ± 0.189% for38Ar/36Ar and 40Ar/36Ar ratios, respectively. These values were determined by mea-suring referenced and cross-validated laboratory standards including an established atmospheric air standard (Lake Erie, Ohio Air), the Yel-lowstone MM standard, and a series of synthetic natural gas standards obtained from Praxair including known and cross-validated con-centrations of C1to C5hydrocarbons, N2, CO2, O2, and each of the noble gases (Harkness et al., 2017;Moore et al., 2018).
5. Results
5.1. Chemical and isotopic (δ18O-H
2O, δD-H2O, δ13C-TDIC, δ34S-SO4, and
87Sr/86Sr) composition of waters
Chemical compositions of the sampled waters are reported in Table 1, where water temperature (from 8.7 to 26°C), pH (from 5.37 to 8.70), and electric conductivity (from 0.16 to 8.8 mS/cm) are also re-ported. The electroneutrality parameter, calculated according Appelo and Postma (1993), was always < 5% (Table 1).
An initial assessment of the chemical composition of the CCVF thermal and cold waters is obtained by considering the Cl−-SO
42-
-HCO3-(Fig. 3a) and (Na++K+)-Ca2+-Mg2+(Fig. 3b) ternary diagrams (expressed as % meq/L). Most waters show a HCO3-Mg and HCO3-Na composition and rare SO4(Cl)-Ca(Mg) facies. Setting aside #FAG5, #H8, #CRE12, and #CRE15, the CCVF waters have a SO42−/Cl−ratio of ≤1 (Fig. 3a), while that of (Na++K+)/Mg2+is < 8 with the ex-ception of #FD10 (45) (Fig. 3b). In the HCO3−vs. SO42-+Cl−diagram (Fig. 4), in which iso-TIS (Total Ionic Salinity) lines are drawn, most waters are between 3 and 50 meq/L, whereas #H6, #CRE12, #H1, and #H7 have a higher TIS (71, 85, 95, and 184 meq/L, respectively). In detail, 3 groups of waters with different compositions can be re-cognized, as follows:
i) HCO3-Mg(Ca): this group includes 11 water samples (#BE1, #H3b,
#BDS4, #CH7, #H5, #FO8, #JA1, #VC6, #FAP2, #H2, and #FLT9), characterized by TIS ranging from 2.86 (#FLT9) to 44.57 (#CH7) meq/L, pH values from 5.6 (#FAP2) to 6.5 (#CH7), and temperatures from 13 (#BDS4) to 22°C (#H3b);
ii) HCO3-Na: 10 water samples (#H4, #FO1, #VP1, #H6, #BLC1, #BLC2, #BLC3, #H7, and #H1) are referred to this group. They have with relatively high TIS, being comprised between 21.88 (#VP1) and 184.04 (#H7) meq/L, slightly acidic pH (5.9–6.2), and temperatures from 17 (#FO1) to 26°C (#H7);
iii) SO4-Cl-Ca(Mg): this group consists of 6 water samples with both low (#FD10, TIS = 2.94 meq/L, pH = 5.4, t = 17°C) and inter-mediate salinity (#UBA1, #FAG5, #H8, #CRE12, and #CRE15, TIS = 21.48 ÷ 84.58 meq/L), moderately neutral to basic pH va-lues (up to 8.7), and temperatures ranging from 8.7 to 22°C. Fluoride and Br− contents were generally < 0.1 mg/L, whereas NH4+and NO3−had a large variability, ranging from < 0.1 to 10.6 (#CRE12) mg/L and from < 0.1 to 101 (#FAG5) mg/L, respectively.
Among the measured trace elements (Table 2), the highest con-centrations were measured for Fe (up to 13,100 μg/L: #FO1), Sr (up to 9765 μg/L: #CRE12), Li (up to 1676 μg/L: #FO1), B (up to 1009 μg/L: #H1), and Mn (up to 784 μg/L: #BE1) (Fig. 1S a, Supplementary Ma-terial) whereas, setting aside Ba, Al, and Rb (up to 224, 157, and 98 μg/ L measured for #BCL1, #FD10, and #H1, respectively), all the other elements were < 50 μg/L (Fig. 1S b, Supplementary Material) and often characterized by contents below the instrumental detection limit.
The isotopic composition of oxygen (δ18O), hydrogen (δD), and strontium (87Sr/86Sr) in water, carbon in the TDIC (δ13C–TDIC), and sulfur in SO4 (δ34S-SO4) are reported in Table 1. The δ18O, δD, and δ13C–TDIC values measured in 19 samples (#VP1, #FAP2, #BDS4, #FAG5, #VC6, #CH7, #FO8, #H1, #UBA1, #FO1, #JA1, #BE1, #FD10, #FLT9, #CRE12, #CRE15, #BLC1, #BLC2, and #BLC3), ranged from −8.58 to −0.68‰ and from −57.7 to −18.9‰ vs. VSMOW, and from −6.80 to +2.21‰ vs. VPDB, respectively. The 87Sr/86Sr isotopic ratios measured in six samples (#H1, #H2, #JA1, #FLT9, #CRE15, and #BLC1) ranged between 0.70415 (#JA1) and 0.71623 (#FTL9). The δ34S-SO
4 values, analyzed in seven samples (#H1, #JA1, #FLT9, #CRE12, #CRE15, #BLC1, and #BLC3), were between 10.7 (#JA1) and 18.3 (#BLC1) ‰ vs. V-CDT.
5.2. Chemical and isotopic (δ13C-CO
2) composition of dissolved gases
Dissolved gases were determined in springs (#H2), rivers (#H8), bubbling pools (#H3b), and upwelling groundwater (#BLC1 and #BLC3). The chemical composition (in mmol/L) and the carbon iso-topes in CO2of the dissolved gases for the CCVF waters are listed in Table 3. CO2 was by far the most abundant gas species (up to 18.5 mmol/L), followed by N2(up to 0.57 mmol/L). Methane, O2, and He were up to 0.0039, 0.16, and 0.00023 mmol/L, respectively, while H2was mostly < 0.00001 mmol/L. The δ13CCO2values in the dissolved gases were measured in the #H2 (−4.1‰ vs. VPDB) and #H3B (−3.7‰ vs. VPDB) samples (Table 3).
5.3. Chemical and isotopic (δ13C
CO2, R/RAand40Ar/36Ar) compositions
of free gases
The chemical (CO2, H2S, N2, CH4, Ar, O2, H2, Ne, and He) and isotopic (carbon in CO2, He, and Ar) compositions of the free-gas samples from CCVF (#H1, #H3, #H4, #H5, #H6, #H7, #VP1, #FAP2, #LS3, #BDS4, #JA1, and #FO8) are reported inTable 4. Carbon di-oxide (up to 992 mmol/mol) largely dominated the free-gas discharges, followed by N2(from 0.71 to 13.5 mmol/mol), H2S (up to 0.8 mmol/ mol), Ar (from 1 × 10−2to 3.1 × 10−1mmol/mol), and CH
4 (from 5 × 10−4 to 8 × 10−2mmol/mol). Minor concentrations were mea-sured for He, H2, and Ne: up 2.5 × 10−2, 1.4 × 10−3, and 1.5 × 10−4mmol/mol, respectively. No CO was detected.
Table 1 Chemical and isotopic composition of the CCVF waters. Location, sampling date, ID, geographical coordinates, altitude (meter), temperature (°C), pH, electric conductivity (mS/cm), chemical composition, TDIC (Total Dissolved Inorganic Carbon in mol/L), pCO 2 (bar), δ 13C-DIC, δD-H 2 O, δ 18O-H 2 O, δ 34S-SO 4 , 87Sr/ 86Sr of the CCVF waters; n.d.: not determined. The sum of cations (Σ cat) and anions (Σ an) are in meq/L. All the other values are in mg/L. Locality Sampling date ID Type Fuse North East Alt Temp. pH E.C. HCO 3 Cl SO4 NO 3 Br UTM UTM m °C mS/cm mg/L mg/L mg/L mg/L mg/L Cañada Real July 09 H1 Bubbling pool with gas 30 S 4300949 425681 611 21.0 6.23 3.76 2074 269 279 0.1 1.07 Fuente Gallega July 09 H2 Spring 30 S 4299477 425518 609 20.0 5.95 0.16 71 5 4 13.5 < 0.1 El Chorillo July 09 H3 Bubbling pool with gas 30 S 4298798 425806 616 n.d. 5.92 0.92 n.d. n.d. n.d. n.d. n.d. El Chorillo July 09 H3b Bubbling pool with gas 30 S 4298752 425819 585 21.9 6.06 n.d. 750 80 110 9.6 0.39 El Baño Chico July 09 H4 Spring with gas 30 S 4295187 428702 619 n.d. 6.13 2.48 945 197 258 0.1 0.82 Los Baños de Villa Franca July 09 H5 Spring with gas 30 S 4299820 421663 640 n.d. 5.84 n.d. 531 55 54 0.3 0.15 Balneares Cervantes July 09 H6 Well with gas 30 S 4279033 462230 685 17.5 5.92 3.32 1220 327 323 9.0 1.49 Baño del Trujillo July 09 H7 Bubbling pool with gas 30 S 4323862 420979 620 26.0 6.36 8.80 2150 1225 1142 0.9 2.07 Rio Guadiana July 09 H8 River 30 S 4324066 419503 603 22.0 8.7 (*) 2.90 122 457 807 0.5 1.89 Los Bagños Hervideros Villar del Pozo (Villar del Pozo) May 10 VP1 Bubbling pool with gas 30 S 4299778 416075 n.d. 24.6 5.93 1.52 485 58 67 2.1 0.19 Fuente Agria Puertollano (Puertollano) May 10 FAP2 Spring with gas 30 S 4282698 403709 n.d. 19.0 5.62 1.20 677 42 13 0.2 0.19 Los Bagños de La Sacristanía o de Valverde (Calzada de Calatrava) May 10 BDS4 Bubbling pool with gas 30 S 4282250 430575 n.d. 13.0 6.30 1.40 689 82 60 0.1 0.34 Fuente Agria de Granátula (Fuente del Bombo) (Granátula de Calatrava) May 10 FAG5 Spring 30 S 4295083 435340 n.d. 14.3 5.80 1.36 242 147 41 101.0 0.63 La Fuente Agria Valenzuela (Valenzuela de Calatrava) May 10 VC6 Spring 30 S 4300620 432954 n.d. 19.3 5.88 1.51 775 41 24 63.0 0.19 El Chorro May 10 CH7 Spring 30 S 4295991 440857 n.d. 16.2 6.54 2.08 1052 117 83 36.0 0.56 Fontecha 2 (Hervidero del Bagño del Chico) (Aldea del Rey) May 10 FO8 Bubbling pool with gas 30 S 4294473 429330 n.d. 17.0 5.88 1.44 822 47 58 0.2 0.21 Cañada Real May 12 H1 Bubbling pool with gas 30 S 4300949 425681 n.d. 19.9 6.13 n.d. 1982 292 329 0.9 < 0.1 Pozuelo de Calatrava May 12 UBA1 Spring 30 S 4305259 426936 n.d. 8.7 7.17 n.d. 274 224 254 99 0.11 Fontecha 1 May 12 FO1 Spring 30 S 4295187 428700 623 17.3 5.89 n.d. 887 175 245 < 0.1 < 0.1 Jabalon May 12 JA1 River with gas 30 S 4295989 427578 n.d. 16.0 6.01 1.60 967 39 55 < 0.1 < d.l Barranco Grande May 12 BE1 Spring 30 S 4294557 429166 614 15.0 5.96 1.53 915 103 145 < 0.1 < d.l Fuencaliente (Fuente del Compadre) May 12 FD10 Spring 30 S 4251371 385817 718 17.5 5.37 n.d. 16 19 21 11.86 < d.l Fuencaliente (Fuente La Teja) May 12 FLT9 Spring 30 S 4250969 386141 641 13.9 5.90 n.d. 51 10 14 0.01 < d.l Rio Guadiana May 12 CRE12 River 30 S 4324858 425773 600 14.4 7.36 n.d. 305 272 1361 4.96 0.54 Rio Jabalon (upstream) May 12 CRE15 River 30 S 4283587 510085 864 13.5 7.83 n.d. 275 71 559 9.34 < 0.1 Bolaños de Calatrava July 12 BLC1 Groundwater 30 S 4310691 438977 n.d. n.d. n.d. n.d. 769 253 260 1.11 0.90 Bolaños de Calatrava July 12 BLC2 Groundwater 30 S 4310691 438977 n.d. n.d. n.d. n.d. 866 284 347 0.70 0.88 Bolaños de Calatrava July 12 BLC3 Groundwater 30 S 4310691 438977 n.d. n.d. n.d. n.d. 894 274 339 0.17 0.85 (continued on next page )
Table 1 (continued ) Locality F Na K Ca Mg NH 4 Σcat Σan Err. TDIC pCO 2 δ 13C-DIC δD-H 2 O δ 18O-H 2 O δ 34S 87Sr/ 86Sr mg/L mg/L mg/L mg/L mg/L mg/L meq/L meq/L % mol/L bar ‰ V-PDB ‰ V-SMOW ‰ V-SMOW ‰ V-CDT Cañada Real 0.21 541 77 45 235 1.03 46.55 47.40 −0.91 0.0710 0.961 −0.25 −59.7 −9.12 n.d. 0.70697 Fuente Gallega 0.23 9 4 9 7 < 0.01 1.52 1.62 −3.24 0.0041 0.074 −6.31 −58.0 −8.85 n.d. 0.70757 El Chorillo n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. −2.82 −56.0 −8.16 n.d. n.d. El Chorillo 0.34 115 26 65 86 0.10 15.80 17.00 −3.66 0.0332 0.560 −2.26 −57.7 −8.90 n.d. n.d. El Baño Chico 0.66 335 52 58 117 7.60 28.16 26.42 3.18 0.0380 0.523 −1.49 −61.2 −9.80 n.d. n.d. Los Baños de Villa Franca 0.30 85 13 46 68 1.93 11.77 11.39 1.66 0.0346 0.602 −2.95 −54.5 −8.13 n.d. n.d. Balneares Cervantes 0.23 333 13 168 142 0.06 34.57 36.10 −2.16 0.0675 1.100 −2.00 −50.6 −7.04 n.d. n.d. Baño del Trujillo 0.10 950 117 193 456 0.77 90.43 93.61 −1.72 0.0595 0.736 −0.27 −61.8 −9.28 n.d. n.d. Rio Guadiana 0.31 234 23 172 166 n.d. 32.64 31.81 1.29 0.0020 0.000 n.d. n.d. n.d. n.d. n.d. Los Bagños Hervideros Villar del Pozo (Villar del Pozo) 0.41 112 10 31 51 0.04 10.86 11.02 −0.70 0.0261 0.527 −1.25 −54.3 −7.73 n.d. n.d. Fuente Agria Puertollano (Puertollano) 0.31 47 9 71 79 0.45 12.32 12.56 −0.95 0.0662 1.340 −3.98 −55.7 −8.58 n.d. n.d. Los Bagños de La Sacristanía o de Valverde (Calzada de Calatrava) 0.32 45 17 58 109 0.04 14.26 14.86 −2.05 0.0244 0.258 −1.36 −55.3 −8.44 n.d. n.d. Fuente Agria de Granátula (Fuente del Bombo) (Granátula de Calatrava) 0.35 33 6 120 40 0.02 10.88 10.60 1.33 0.0182 0.294 −1.15 −51.5 −7.51 n.d. n.d. La Fuente Agria Valenzuela (Valenzuela de Calatrava) 0.31 40 13 129 87 0.02 15.68 15.38 0.98 0.0464 0.827 −0.93 −50.9 −7.14 n.d. n.d. El Chorro 0.38 48 15 189 119 0.03 21.72 22.86 −2.56 0.0275 0.227 −1.45 −53.2 −7.55 n.d. n.d. Fontecha 2 (Hervidero del Bagño del Chico) (Aldea del Rey) 0.45 45 28 57 115 0.26 14.99 16.01 −3.30 0.0513 0.859 −1.08 −57.7 −8.50 n.d. n.d. Cañada Real 3.29 456 28 156 230 n.d. 47.27 47.60 −0.34 0.0770 1.120 −4.78 −45.4 −5.32 14.1 n.d. Pozuelo de Calatrava 1.06 32 16 130 95 n.d. 16.13 17.69 −4.60 0.0052 0.013 −5.32 −46.5 −5.70 n.d. n.d. Fontecha 1 0.62 286 26 54 130 n.d. 26.55 24.59 3.83 0.0531 0.886 −6.23 −45.1 −7.12 n.d. n.d. Jabalon 0.29 127 28 117 92 n.d. 19.65 18.09 4.14 0.0487 0.722 −4.93 −46.1 −5.44 10.7 0.70415 Barranco Grande < d.l 159 38 75 116 n.d. 21.16 20.92 0.57 0.0506 0.757 −6.62 −44.9 −7.06 n.d. n.d. Fuencaliente (Fuente del Compadre) 0.22 8 7 17 2 n.d. 1.50 1.45 1.71 0.0029 0.062 −6.80 −44.6 −7.00 n.d. n.d. Fuencaliente (Fuente La Teja) < d.l 8 1 12 6 n.d. 1.46 1.41 1.72 0.0035 0.054 −4.90 −46.3 −5.72 14.7 0.71623 Rio Guadiana 4.28 144 16 459 168 10.66 43.47 41.11 2.79 0.0054 0.009 2.21 −18.9 −0.68 16.3 n.d. Rio Jabalon (upstream) < d.l 16 10 246 82 n.d. 19.96 18.29 4.38 0.0046 0.003 1.51 −20.1 −1.32 13.3 0.70872 Bolaños de Calatrava 0.61 183 29 117 96 4.3 22.63 24.97 −4.92 0.0150 0.575 −3.40 −45.7 −4.80 18.3 0.70794 Bolaños de Calatrava 0.55 201 61 143 107 15 27.05 29.45 −4.23 0.0177 0.638 −2.80 −43.2 −3.90 n.d. n.d. Bolaños de Calatrava 0.53 198 61 147 105 18 27.15 29.44 −4.04 0.0185 0.658 −2.70 −43.6 −3.90 15.9 n.d. (*) computed CO32-= 2.8 mg/L.
Among volatile organic compounds (VOCs), the most abundant species were related to light alkanes (C2H6and C3H8: up to 3 × 10−3 and 1 × 10−4mmol/mol, respectively), while C
6H6was found at very low concentrations (up to 8.4 × 10−5mmol/mol); i-C
4H8and i-C4H10 were detected in six samples (#H1, #H3, #H4, #H5, #H6, #H7) at concentrations up to 9 × 10−6and 8 × 10−6mmol/mol, respectively.
The δ13C values in CO
2in the free gas samples ranged from −6.8 (#FAP2) to −3.2 (#LS3) ‰ VPDB. The helium isotopic ratios (mea-sured in #H4, #H5, #H6, #VP1, and #FO8), corrected for the presence of air in the mixture by using the He/Ne ratio, varied from 1.06 (#H6) to 2.73 (#VP1) R/RA.Peréz et al. (1996)(in two unspecified sites from CCVF) measured 1.81 and 1.82 R/RA, respectively, i.e. in the range of those determined in the present study, though lower. The Ar isotopic ratios (40Ar/36Ar; measured in the gas samples where the helium iso-topes were analyzed) exceeded that of the air (295.5) since they were ranging from 300 to 407.
6. Discussion
6.1. Processes governing the chemical and isotopic composition of waters
The δ18O-H
2O and δD-H2O values of most waters are distributed along the Global Meteoric Water Line (GMWL;Craig, 1961) (from −9.8 to −7.0‰ and from −61.8 to −44.5‰ for δ18O-H
2O and δD-H2O, respectively), pointing to a meteoric origin (Fig. 5). The relatively large isotopic variations can be explained in terms of different altitudes of the recharge areas. The CCVF thermal and cold waters are fed by rainfall between 600 and 700 m above sea level in according to the relationship between δ18O vs. altitude of about −0.3‰/100 m, (e.g.Cruz-San et al., 1992). A few samples divert from the GMWL, possibly reflecting either prolonged water-rock interactions (though not supported by the rela-tively low salinity) or evaporation processes, and may play an im-portant role due to the peculiarity of the fluid discharges since the waters often discharge inside small pools with scarce water exchange (Fig. 2). This is supported by the fact that i) the evaporation line re-ported inFig. 5intercepts the GMWL at the isotopic values of #FAP2, #FO8, and #BDS4; and ii) the upwelling groundwater samples (#BLC1, #BLC2, #BLC3; up to −3.9‰ and −43.2‰ for δ18O-H
2O and δD-H2O, respectively) are distributed along the same evaporation line. A second evaporation line consists of the river waters (#CRE15, #CRE12 up to −0.7‰ and −18.2‰ for δ18O-H
2O and δD-H2O, respectively). Ac-cording toGonfiantini (1986), the slope (“s” in Fig. 5) of these two evaporation lines is of 3.9 and 5, respectively, which corresponds to an area with relatively low humidity (h ≤ 0.5), as expected in this part of Spain where temperatures during the day can reach ≥30°C with low humidity.
As previously stated, three groups of waters (HCO3-Mg(Ca), HCO3 -Na, and rare SO4-Cl-Ca(Mg)) were recognized. The CCVF waters can mainly be regarded as the result of meteoric waters that, while in-filtrating underground, react with (i) volcanic products characterized by Mg-rich volcanic alkaline rocks; (ii) silicate rocks (mainly re-presented by Paleozoic quartzites and sandstones); (iii) a deep-sourced CO2-rich gas phase; and (iv) waters from the Jabalón and Gaudiana rivers, which are rich in solutes derived by interactions with the Triassic evaporitic rocks (e.g. the Buntsandstein-Muschelkalk-Keuper Formations inFig. 1).
The dissolution of CO2causes a decrease in the pH (down to 5.3), enhancing the dissolution of Mg(Ca)-rich carbonate minerals that, consequently, increase the abundances of Ca, Mg, and HCO3 in the water. Accordingly, most samples from the study area are approaching the stoichiometric (Ca + Mg)/HCO3 ratio (Fig. 2S, Supplementary Material). The Ca*3-HCO3-Mg*10 triangular plot is reported inFig. 3S (Supplementary Material), where the water composition expected for the CO2-driven dissolution of the dominant mineral phases that char-acterize the CCVF lithologies (Fig. 1) by stoichiometric reactions (e.g. calcite, dolomite, diopside, anorthite, and fosterite; eqs.(2)–(6)) are
Fig. 3. Geochemical classification of the CCVF waters by (a) SO4-Cl-HCO3and (b) Ca-(Na + K)-Mg triangular diagrams.
Fig. 4. (SO4+Cl) vs. HCO3(in meq/L) binary diagrams of the CCVF waters. The dashed lines represent iso-Total Ionic Salinity (TIS) values.
reported as follows: + + = + + CaCO CO H O Ca 2HCO . (calcite) 3 2 2 2 3 (2) + + = ++ ++
CaMg(CO ) 2CO 2H O Ca Mg 4HCO .
(dolomite)
3 2 2 2 2 2 3
(3)
+ + = ++ ++ +
CaMgSi O 4CO 2H O Ca Mg 4 HCO 2SiO .
(diopside)
2 6 2 2 2 2 3 2
(4)
+ + = + ++
CaAl Si O 2CO 3H O Al Si O (OH) Ca 2HCO . (anorthite)
2 2 8 2 2 2 2 5 4 2 3
(5)
+ + = ++ +
Mg SiO 4CO 2H O 2Mg 4HCO SiO . (forsterite)
2 4 2 2 2 3 2
(6) A Ca(Mg)-HCO3 composition is attained in the initial stages of water-rock interaction processes due to calcite dissolution even in small amounts. This is related to the dissolution rate of calcite, being 2 to 6 orders of magnitude higher than that of silicate minerals, depending on the pH (Stumm and Morgan, 1996and references therein). Setting aside samples #CRE15, #CRE12, #H8, #UBA1, #FD10, and #FAG5 (Fig. 3S, Supplementary Material), most waters have a HCO3-Mg(Ca) and HCO3 -Na composition likely due to water-rock interactions with a CO2-rich gas phase and silicate minerals (eqs.(1)–(6)).
The chemical characteristics of the HCO3-Mg(Ca) waters (TIS = 3 ÷ 44 meq/L) are typical of the first stages of interaction between meteoric and rocks, where the dominant concentration of Mg is attained by the
alkaline rocks. The peculiar HCO3-Na composition, characterized by a clear enrichment in Na with respect to the stoichiometric Na/Cl ratios (Fig. 6), is likely due to incongruent dissolution processes affecting Na-silicates during relatively long-lasting water-rock interaction as in-dicated by the TIS values (TIS = 22 ÷ 184 meq/L), favored by the presence of a CO2-rich gas phase. The SO4-Cl-Ca(Mg) waters (#CRE15, #CRE12 #H8, #FD10, #FAG5, and #UBA1) are likely influenced by an evaporitic component represented by the Triassic Formation drained by Guadiana and Jabalón rivers before entering the study area (Fig. 1). This implies that the waters discharging in the CCVF may be mixed with a SO4-Ca-rich shallow aquifer related to the two rivers. This is parti-cularly evident when samples #UBA1, #FD10, and #FAG5 springs are considered since the sulfur isotopic composition (+10.7 to +18.3‰ vs. CDT;Table 1) intimately resembles that of gypsum/anhydrite from the Triassic evaporitic rocks (12.5–16.6‰ vs. CDT;Ortí et al., 2014) in the Beltic Cordillera of the Germanic-type facies (Buntsandstein, Mu-schelkalk, and Keuper) (López-Gómez et al., 1993). Evaporitic rocks undergo water-rock interactions more rapidly than silicate rocks, hence their solutes have a major influence on water chemistry even if they are sporadically outcropping (Meybeck, 1987). The Ca/Na, Mg/Na, and SO4/Na molar ratios are particularly suited to distinguish waters in-teracting within the CCVF rocks, as shown inFig. 7a and b. The three groups of waters are related to mixing processes with the SO4(Ca,Mg) evaporitic component that interacts with the shallow waters, the latter being due to the interaction with the alkaline volcanic rocks, whereas those with a HCO3-Na composition seem to be affected only to a very minor extent. Consequently, the relatively low Mg/Ca and Ca/Na ratios (mean value: 0.44 and 0.22, respectively) recorded by HCO3-Na waters
Table 2
–Trace element concentrations (in μg/L) in selected waters from CCVF.
ID Al As Ba B Cs Co Cu Fe Hg Li Mn Ni Rb Se Sr Zn μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L H1* 4.7 0.5 22 1009 19 5.7 0.1 120 0.7 1203 311 14.9 98 < 0.1 94 26 H2 4.9 0.1 81 129 < 1 0.1 3.1 7 1.1 82 2 2.1 5.4 < 0.1 40 9.0 H3 34 0.1 39 423 5.0 1.1 2.7 8 11 594 127 6.7 40 0.1 109 18 FAP2 19 3.8 46 90 6.0 2.1 0.1 6683 0.7 160 509 5.0 19 < 0.1 22 9.3 BDS4 2.6 0.9 40 105 17 1.3 0.4 87 0.8 588 150 7.7 41.4 < 0.1 22 11 FAG5 33 0.1 93 45 < 1 0.6 5.9 8 0.8 64 31 6.4 9.2 0.9 391 23 VC6 23 0.5 199 105 < 1 0.3 21 < 5 1.0 61 6.8 3.4 5.4 0.2 1062 33 UBA1 1.9 0.5 131 76 < 1 0.2 0.4 6 10 52 2.2 0.7 1.1 0.3 3004 22 FO1 55 8.3 15 798 13 14 0.2 13100 0.3 1676 612 50.8 82 < 0.1 13 50 JA1 3.5 2.9 111 171 < 1 19 0.5 7600 0.4 279 544 2.3 63 < 0.1 455 13 BE1 20 2.3 27 417 13 16 0.1 4377 1.5 1236 784 38.7 61 < 0.1 18 28 FD10 157 4.7 26 34 < 1 3.8 19 23 2.4 5.0 75 2.8 4.1 0.2 24 32 FLT9 1.7 1.3 29 38 < 1 3.4 0.2 4361 0.7 28 481 7.9 2.8 < 0.1 19 12 CRE12 3.4 1.8 23 151 < 1 0.9 0.5 203 0.2 84 71 1.6 4.4 1.0 9765 2.6 CRE15 4.1 1.3 34 220 < 1 0.4 1.2 15 0.4 76 31 0.8 2.2 0.9 4063 4.0 BLC1 3.8 17 224 319 < 1 6.8 1.0 76 0.7 101 284 27.5 15 0.1 1481 4.7 BLC2 32 8.8 82 236 1.0 1.8 1.0 145 1.0 117 431 7.3 30 0.2 1559 5.4 BLC3 26 5.8 79 246 1.0 1.3 0.7 72 0.2 114 377 5.5 32 0.1 1585 7.0 (*) July 09. Table 3
Chemical and isotopic composition for the CCVF dissolved gases. The gas concentrations are in mmol/L.
ID Sampling date CO2 N2 Ar CH4 O2 Ne He H2 δ13C-CO2 N2/Ar
mmol/L mmol/L mmol/L mmol/L mmol/L mmol/L mmol/L mmol/L V-PBD
H2 July 09 3.05 0.570 0.014 0.0002 0.1321 0.00001 0.00004 < 0.00001 −4.1 41.1
H3B July 09 18.48 0.074 0.002 0.0039 0.0233 < 0.00001 0.00023 < 0.00001 −3.7 41.7
H8 July 09 0.15 0.475 0.008 0.0003 0.0033 < 0.00001 0.00001 < 0.00001 n.d. 58.9
BLC1 July 12 0.00 0.380 0.011 < 0.0001 0.1626 0.00001 n.d. < 0.00001 n.d. 36.2
Table 4 Chemical and isotopic composition of the CCVF free gases. The main gases are in mmol/mol whereas C2 H6 ,C3 H8 ,i-C 4 H10 ,i-C 4 H8 ,and C6 H6 are in μmol/mol. ID CO 2 H2 S N2 CH4 Ar O2 Ne H2 He C2 H6 mmol/mol mmol/mol mmol/mol mmol/mol mmol/mol mmol/mol mmol/mol mmol/mol mmol/mol μmol/mol H1* 989.48 0.06 9.85 0.01 0.14 0.46 0.00007 0.00051 0.00050 0.428 H3 985.26 0.08 11.85 0.03 0.29 2.46 0.00015 0.00113 0.02546 3.204 H4 998.94 0.05 0.99 0.01 0.01 0.00 0.00000 0.00002 0.00017 0.142 H5 998.98 0.07 0.71 0.00 0.02 0.22 0.00001 0.00003 0.00117 0.014 H6 984.71 0.06 13.50 0.08 0.31 1.33 0.00018 0.00062 0.00119 1.584 H7 992.60 0.05 5.14 0.01 0.11 2.09 0.00006 0.00139 0.00041 0.404 VP1 990.52 0.03 8.46 0.01 0.14 0.85 0.00007 0.00022 0.00035 0.036 FAP2 991.50 0.03 7.75 0.02 0.22 0.48 0.00012 0.00006 0.00148 0.005 LS3 988.91 0.06 9.65 0.01 0.19 1.18 0.00010 0.00026 0.00184 0.056 BDS4 991.82 0.07 6.95 0.02 0.15 0.99 0.00008 0.00005 0.00245 0.017 FO8 988.17 0.02 9.95 0.01 0.20 1.65 0.00011 0.00009 0.00316 0.011 H1** 992.52 < d.l. 6.70 0.01 0.10 0.67 0.00006 0.00011 0.00085 0.021 JA1 989.73 < d.l. 8.95 0.01 0.13 1.17 0.00007 0.00009 0.00111 0.026 ID C3 H8 i-C4 H10 i-C4 H8 C6 H6 δ 13C-CO 2 R/R A N2 /Ar He/Ne 36Ar/ 40Ar CO2 / 3He (x10 11) μmol/mol μmol/mol μmol/mol μmol/mol ‰ V-PBD H1* 0.0188 0.0052 0.0031 0.0616 −3.7 n.d. 71 7 n.d. n.d. H3 0.1010 0.0000 0.0088 0.0746 −3.7 n.d. 41 172 n.d. n.d. H4 0.0059 0.0003 0.0009 0.0035 −4.5 2.40 135 40 353 0.2 H5 0.0010 0.0000 0.0001 0.0015 −4.2 2.16 40 129 300 2.8 H6 0.0594 0.0076 0.0046 0.0503 −4.9 1.06 43 7 407 5.6 H7 0.0184 0.0059 0.0044 0.0844 −5.2 n.d. 46 7 n.d. n.d. VP1 0.0011 n.d. n.d. 0.0215 −4.1 2.73 63 5 312 7.5 FAP2 0.0005 n.d. n.d. 0.0015 −6.8 n.d. 36 13 n.d. n.d. LS3 0.0015 n.d. n.d. 0.0164 −3.2 n.d. 51 19 n.d. n.d. BDS4 0.0017 n.d. n.d. 0.0085 −5.8 n.d. 46 29 n.d. n.d. FO8 0.0008 n.d. n.d. 0.0026 −4.4 2.06 50 29 310 1.1 H1** 0.0011 n.d. n.d. 0.0132 −4.5 n.d. 66 15 n.d. n.d. JA1 0.0012 n.d. n.d. 0.0111 −4.8 n.d. 68 15 n.d. n.d. (*) July 09. (**) May 12.
(associated with a CO2-rich gas phase), are likely referring to the dis-solution of silicate minerals, which characterize the Paleozoic basement rocks. The HCO3-Mg(Ca) waters have higher Mg/Ca and Ca/Na ratios (mean value: 1.38 and 0.85, respectively) than those classified as HCO3 -Na (Fig. 7a and b). The SO4-Cl-Ca(Mg) waters have higher SO4/Na (=2.36) and Ca/Na (=2.77) molar ratios, suggesting a strong con-tribution of SO4, Ca, and Na by an evaporitic component.
The relatively high concentrations of Sr (up to 9765 μg/L) observed in samples #VC6, #BA1, #CRE12, #CRE15, #BLC1, #BLC2, and #BLC3 (Table 2) corroborate the occurrence of water-rock interaction processes, involving carbonate and sulfate minerals, since Sr is the main substitute for divalent ions, such as Ca, and is easily released during dissolution. The Sr isotopic composition is a useful parameter that can be used to constrain water-rock interaction processes (e.g.Blum et al., 1994;Négrel et al., 2001;Nisi et al., 2008). The87Sr/86Sr ratio is as-sociated with the age of the geologic formations. If Rb is not present in the mineral phase, the87Sr/86Sr ratio records the isotopic ratio of the medium from which a certain mineral was precipitated (e.g. calcite and gypsum/anhydrite). Since natural processes do not fractionate Sr iso-topes, the measured differences in the87Sr/86Sr ratios are likely due to either the isotopic signature inherited by the rocks with which the waters are interacting or mixing processes among rocks of different ages. The87Sr/86Sr vs. HCO
3/HCO3+SO4ratios ofFig. 8 reports the selected waters (#FLT9, #CRE15, #BLC1, #H1, #H2, and #JA1) where the87Sr/86Sr ratios were measured and those of the main CCVF rock end-members occurring in the study area, i.e., Triassic evaporites (Ortì at al., 2014), basaltic rocks (Ancochea and Moro, 1981), and Paleozoic rocks from the Betic belt (Benito et al., 1999). The87Sr/86Sr isotopic ratios of the CCVF waters are varying between the relatively high radiogenic values of the Paleozoic basement (#FLT9, 87Sr/86Sr = 0.71623) and those of the volcanic rocks, which are char-acterized by the low Sr isotopic ratio (e.g.,Cebria et al., 1995) and are similar to that recorded in sample #JA1 (0.70415). Setting aside the river water of #CRE15 [HCO3−/(HCO3−+ SO42) = 0.23], the rela-tively low variability of the HCO3−/(HCO3−+ SO42−) ratios (from 0.70 to 0.93) corresponds to a high variability in terms of 87Sr/86Sr ratios (from 0.70415 to 0.71623). The highest HCO3−/ (HCO3−+ SO42−) ratios are found in #JA1, #H1, and #H2 and related to the dissolution of a CO2-rich gas phase (Table 3). A similar process may have affected the samples #FLT9 and #BLC1, although a con-tribution by Paleozoic and Triassic evaporitic rocks, respectively, is expected. The lowest HCO3−/(HCO3− + SO42−) ratio was found in #CRE15 (0.23), which is characterized by a dominant SO4-Ca facies (Ca/SO4molar ratios ≈1) and has inherited the Sr isotopic value by the dissolution of the gypsum/anhydrite from the Triassic evaporitic layers. As previously mentioned, large areas of the river basin (#CRE15; δ34 S-SO4= +13.3‰ vs. CTD) are interacting with this lithology (Fig. 1).
Fig. 5. δD vs. δ18O binary diagrams for the CCVF waters. Global Meteoric Water Line (GMWL) byCraig (1961); s is the slope at different relative humidities calculated according toGonfiantini (1986).
Fig. 6. (a) Na vs. Cl (in meq/L) and (b) SO4vs. Cl (in meq/L) binary diagrams for the CCVF waters that include the stoichiometric (1:1) and seawater (1:1.2) and seawater (1:1.2) line, respectively.
Field evidence, like bubbling waters and sporadic gas blowouts, suggests that the composition of these waters is related to the influx of a deep CO2-rich gas that is transported to shallow aquifers, i.e., the aquifer hosted in the basaltic volcanic rocks, though partly affected by those of the Guadiana and Jabalón rivers.
The relatively high concentrations of Co (#JA1, #FO1, and #BE1), Fe (#FO1, #BE1, #FD10, #FLT9, #FAP2, and #JA1) and Mn (#FO1, #BE1, #FAP2, and #JA1) are likely related to their relatively low pH (from 5.37 to 6.06), i.e. the lowest pH values measured in the CCVF waters. These features are likely related to the hydrothermal activity, which was either subsequent to or associated with the magmatic ac-tivity, as suggested by the presence of deposits of Mn-Fe oxide with relatively high Co contents (Higueras and Millá, 2011). The relatively high concentrations of Li (#H1, #FO1, and #BE1) and B (#H1, #FO1) further support the contribution by dissolution of evaporitic rocks.
6.2. Comparison between theoretical and observed compositions: Total
Dissolved Inorganic Carbon (TDIC) and δ13C-TDIC isotopic ratios
The range of variation of the PCO2in waters is related to the content of dissolved carbonic acid (H2CO3) that is the main acidic substance driving mineral dissolution reactions. The higher the dissolved CO2, the lower the initial groundwater pH. The low pH is then buffered by mi-neral weathering in soils and outcropping bedrocks. Calcite dissolution is the most common and effective buffering reaction. The alteration of silicate minerals also consumes [H]+, although these reactions proceed more slowly. Carbonate dissolution is sensitive to the partial pressure of CO2, i.e. the higher the dissolved CO2the greater the solubility of cal-cite. This produces a relatively high TDIC content. Weathering of sili-cate minerals has a different effect on the carbonate system. The TDIC is solely derived from the CO2consumed by alteration of feldspars, e.g. anorthite to form kaolinite (eq.(5)). As a consequence, in such reac-tions, the only change in the carbonate system is the associated increase in pH, which shifts the distribution of TDIC species to HCO3(the main species for pH values between pH 6.4 and 10.3).
Mineral dissolution processes affecting most CCVF waters appear to be strictly controlled by PCO2. In the PCO2 vs. pH binary diagram of Fig. 4S (Supplementary Material), the studied waters are compared
with the theoretical curves representing three iso-TDIC concentrations lines (= 10, 100, and 1000 mgHCO3/L, respectively). The HCO3-Mg (Ca) and HCO3-Na waters show high TDIC and PCO2values (TDIC > 1000 mg HCO3/L and up to 10+0.1bar, respectively) whereas the pH ranges within a relatively narrow interval (from 5.5 to 6.5). The sam-ples #FD10, #FLT9, #H2, #UBA1, #CRE12, #CRE15, and #H8 are distributed along a decreasing trend since the TDIC values become lower as the pH increases (∼100 < TDIC < 500 mgHCO3/L and 10−3.6< P
CO2< 10−1.2; respectively), the latter ranging between 5.4 and 8.7.
During weathering reactions, the carbon isotopic composition of TDIC (δ13C-TDIC) tends to evolve towards higher values (e.g.Clark and Fritz, 1997). For infiltrating meteoric waters,Frondini et al. (2009)and Nisi et al. (2013b)proposed an evolution model based on the addition of biogenic CO2, deeply derived CO2, and the simultaneous equilibrium dissolution of calcite. The TDIC contents vs. the δ13C-TDIC values are reported inFig. 5S (Supplementary Material), along with the theoretical curves simulating the TDIC and δ13C-TDIC evolution. The theoretical curves were computed by means of the EQ3/6 code, starting from low (TDIC = 5.6 × 10−4mol/kg), to intermediate (TDIC = 4.1 × 10−3mol/ kg), to relatively high (TDIC = 2.0 × 10−2mol/kg) TDIC values. In order to investigate the effects of CO2on the TDIC and δ13C-TDIC values, the input of deep CO2was modeled by the addition of 1.0 × 10−2mol of carbon to the infiltrating waters with a δ13C = −4.3‰ (V-PDB) (the mean value of the measured δ13C-CO
2after excluding the most negative values, see Table 4) and calcite δ13C = 0‰ (V-PDB), while the δ13C value of CO2biogenic added to the solution was δ13C = −20‰ (V-PDB), which is comparable with oxidation processes of organic matter and/or root respiration (Nisi et al., 2016, and reference therein) (solid black curve inFig. 5S, Supplementary Material). By comparison, the “deeply derived” CO2 is assumed to be originated from degassing mantle (Chiodini et al., 1999,2000,2004). The inspection ofFig. 5S (Supple-mentary Material) shows that the #FAP2, #H1, #JA1, #FO1, #BE1, #UBA1, #H2, #FD10, and #FLT9 water samples are characterized by 2.9 × 10−3< TDIC < 7.7 10−2mol/kg and tend to be mainly dis-tributed along the theoretical curves, representing the addition of deep CO2(dash black curve inFig. 5S, Supplementary Material), whose iso-topic signature (δ13C-TDIC values from −6.80 to −3.98‰ vs. V-PDB), suggests that the CCVF aquifer system is affected by CO2-rich gases from a deep source. A significant number of the CCVF waters are characterized by TDIC values ranging from 1.8 × 10−2and 6.7 × 10−2mol/kg with those of δ13C-TDIC varying from −2.8 to −0.27‰ (V-PDB), i.e. related to the interaction with a deep-seated carbon. This hypothesis is sup-ported by both the presence of large areas of CO2degassing (Elio et al., 2015) and the PCO2values estimated at reservoir conditions (pressure of 63 bars byGonzález Cárdenas et al., 2015), which is, as expected, much higher than that computed at sampling conditions.
6.3. Origin of gases
The chemical composition of the free dissolved and gases from the CCVF is largely dominated by CO2, as it approaches concentrations up to 18.4 mol/L and 999 mmol/mol respectively (Tables 3 and 4). In the free-gases, N2(up to 9.95 mmol/mol), and Ar (up to 0.31 mmol/mol) are subordinate with respect to CO2. Much lower contents were re-corded for H2S (up to 0.08 mmol/mol), H2(up to 0.0011 mmol/mol), CH4 (up to 0.034 mmol/mol), He (0.025 mmol/mol), and VOC (ΣVOC= 3.38 μmol/mol).
According to the N2/Ar ratios (Tables 3 and 4), these inert gases are mainly related to an atmospheric component, their ratios ranging be-tween that of air (83) and that of ASW (Air Saturated Water; N2/ Ar = 38), with the exception of sample H4 (N2/Ar = 135) for which a contribution from deep-seated N2can be invoked, this sample showing the lowest concentration of Ar and O2and the highest content of CO2. Based on the40Ar/36Ar ratio (from 300 to 407) that often exceeds that of the air, the presence of a radiogenic component of Ar in the studied
Fig. 8. Binary diagram of87Sr/86Sr isotopic ratio vs. HCO
3/(HCO3+SO4) ratio. The strontium isotopic interval for the CCVF basalts (Ancochea and Moro, 1981), the Triassic Keuper facies (Ortì at al., 2014), and the Paleozoic rocks (Benito et al., 1999) are reported for comparison.
samples can be suggested. The predominantly atmospheric origin of N2 and Ar can also be visualized by using the N2-Ar-He triangular diagram (Giggenbach, 1996, Fig. 9). This diagram illustrates how the CCVF gases appear to be related to a mixing process between a mantle component (as suggested by the relatively high R/Ra values,Table 4) and an atmospheric component (either air or ASW Air Saturated Water). According to the R/Ra values measured in the ultramafic xe-noliths from the CCVF (ca. 6.5Ra) hosted in the alkaline rocks,Martelli et al. (2011)suggested that these values can be associated with meta-somatic processes due to ascending HIMU-type astenospheric melts and they are in agreement with what observed in other xenoliths from the European mantle (e.g.Dunai and Bauer, 1995). The helium isotope (R/ Ra) ratios in the CCVF are indeed lower (< 3Ra,Table 4) than those expected for upper mantle volatiles (≈8Ra) and consistent with a mantle component that experienced metasomatism either during the interaction with crustal fluids (Hilton et al., 2002) or, more likely, ty-pical of rift-related continental area (Bell et al., 2013), possibly diluted by a crustal component, associated with the Paleozoic basement. It can thus be estimated that up to 40% of a metasomatized mantle compo-nent is present in the analyzed gases, this value being slightly higher than those calculated in two gas samples collected from the CCVF by Perez et al. (1996). The carbon isotopic values of CO2are between −3.2 and −6.8‰ V-PDB, i.e. in the range suggested for primary mantle carbon (−6.5 ± 2.5‰ V-PDB, e.g. Sano and Marty, 1995). However, the CCVF gas samples show a CO2/3He ratio > 1011, which significantly exceeds that typically measured for strictly mantle-derived gases (2 × 109-1x1010, e.g.Marty et al., 1989;Sano and Marty, 1995; Sano and Williams, 1996) and overlaps or is even higher than those measured in fumarolic discharges associated with subduction-related volcanic systems or typically observed from water washing (Tedesco et al., 2009;Darrah et al., 2013). Consequently, an unequivocal origin for CO2cannot be invoked although the presence of a mantle signature, possibly modified by the input of a shallower (e.g. carbonate) compo-nent as suggested by the CO2/3He ratios, is likely. Nevertheless, sec-ondary processes occurring before the emergence of the CO2-rich bearing thermal waters cannot be ruled out (Venturi et al., 2017). No isotopic data are available for sulfur isotopes in H2S and carbon and hydrogen isotopes in methane, though present at very low concentra-tions, and their origin cannot be unequivocally distinguished. However, if the CH4/(C2H6+C3H8) ratio (the so-called “Bernard parameter”, after Bernard et al., 1978) is considered, all the gas samples are < 1,000, with the exception of #FAG2 (4500), and some of them are close or even < 100. The intermediate values between a thermogenic (“Bernard
parameter” < 100) and a microbial (“Bernard parameter” > 1000) source (Whiticar and Suess, 1990;Jenden et al., 1993) are likely pro-duced by both microbial activity and thermal maturation of sedimen-tary organic matter. However, a thermogenic source can be invoked for samples #H1 to #H7 and #LS3. In addition, the presence of C6H6and C4+ hydrocarbons in the CCVF gases suggests a minor contribution from a thermogenic component (e.g. Tassi et al., 2012). A detailed isotopic investigation on methane and H2S is necessary before for-mulating more specific hypotheses.
6.4. Geothermometry
González Cárdenas et al. (2015) reported that a hydrothermal system below CCVF is present with temperature of 118–120°C. How-ever, no information whether these values were obtained by using li-quid and/or gas geothermometers or derived by exploratory geo-thermal wells were reported. The authors simply referred that the geothermometric data were produced during some projects lead by the Instituto Volcanológico de Canarias (INVOLCAN) and Grupo de In-vestigación: Geomorfología, Territorio y Paisaje en Regiones Volcánicas (GEOVOL). On the other hand,Benítez–Navío and Pulido–Bosch (2010) described the problems related to the application of liquid geother-mometry and suggested a hypothetical equilibrium temperatures of about 70°C. Consequently, in this section, we used our data in order verify whether geothermometric estimations can be applied to the CCVF waters and gases. It is to be pointed out that the outlet tem-peratures of the studied emergences were < 30°C, although a slightly higher temperature was recorded at Baños de Fuencaliente (38°C), was not sampled in present work.
The most commonly used empirical geothermometers are based on the theoretical assumption of equilibrium between water and the ty-pical mineral authigenic assemblage of medium to high temperature (150–300°C) hydrothermal systems (e.g. Giggenbach, 1988; Chiodini et al., 1991). As already discussed, the HCO3-Na waters can be regarded as the deeper liquid component among the studied waters, likely re-leased from and equilibrated within the Paleozoic basement and af-fected by a deeper gas (mostly consisting of CO2) root. The Na-K-Mg1/2 triangular diagram of Giggenbach (1988) (Fig. 6S, Supplementary Material) shows that the HCO3-Na waters are positioned close to the Mg corner (“immature waters”), likely resulting by a mixing process with the shallower aquifer (HCO3-Mg) located inside the alkaline volcanic products. Consequently, it appears that the HCO3-Na deep waters are not the result of high temperature equilibrium. Moreover, the mixing process between upwelling deep waters and a shallow cold aquifer could contribute to the decrease in the concentrations of the deep-re-lated solutes (e.g. SiO2) due to re-equilibration processes. Since no equilibrium is achieved along with the presence of abundant CO2, si-lica, and alkaline(earth) geothermometers (e.g.Fournier and Truesdell, 1973;Fouillac and Michard, 1981;Giggenbach, 1988), when applied to the CCVF waters, give very different and unrealistic deep reservoir temperatures compared to those reported byGonzález Cárdenas et al. (2015). If the equilibrium temperature and physical data described in González Cárdenas et al. (2015)are correct, the CCVF hydrothermal system would occur at a depth of 640 m and a pressure of 63 bars. This would mean that the geothermal gradient at CCVF is comparable to that of Larderello (Italy; e.g.Ceccarelli et al., 1987;Minissale, 1991;Della Vedova et al., 2007). Therefore, unless other evidences to support the data ofGonzález Cárdenas et al. (2015)are provided, such temperature estimations are to be thought overestimated, since they are clearly af-fected by the presence of shallow aquifers.
Gas geothermometry is applied to gases with low solubility (e.g. CH4, H2, and CO2) since the relative ratios are not affected by sig-nificant compositional variations as they rise from the reservoir to the surface. Consequently, the CH4-CO2-H2system (e.g.Giggenbach, 1987; Giggenbach, 1991) was applied to the CCVF gases. However, no rea-listic data were obtained, likely due to the presence of a shallow
Fig. 9. N2/100-Ar-10*He triangular diagram for the free (red diamond) and dissolved (yellow diamond) gases from CCVF. (For interpretation of the refer-ences to colour in this figure legend, the reader is referred to the Web version of this article.)
